Building the Machine: How One Credit Union Is Getting AI Right

Artificial intelligence is often discussed in financial services as a front-end transformation: chatbots, virtual assistants, and member-facing automation. But at Trumark Credit Union, the real shift is happening behind the scenes.

Under Ben Fetterman's leadership, Trumark has transformed digital from a traditional channel into an enterprise-wide driver of growth. More recently, the organization has adopted a thoughtful approach to AI, focusing first on enhancing internal efficiency and strengthening data foundations, while intentionally controlling member-facing solutions through a more conversational and canned AI.

The result is a pragmatic, operationally focused AI strategy that prioritizes scalability, risk management, and member outcomes over speed to market.

Chuck Hayes: How would you describe your current role and Trumark's digital footprint today?

Ben Fetterman: I'm the First Vice President, Digital Solutions Director at Trumark Credit Union. We're a $3.5 billion institution based in Fort Washington, Pennsylvania, serving the greater Philadelphia area.

I've been with Trumark for seven years. I originally joined on the retail side, leading that division. About three years ago, I transitioned fully into a digital-focused role. Since then, my focus has been on building and evolving our digital strategy so that it's not just competitive but differentiated.

Chuck Hayes: You've described a shift in how Trumark thinks about digital. What changed?

Ben Fetterman: Early on, digital was treated like a channel, essentially an alternative to branches or call centers. Online and mobile banking sat within retail.

We challenged that.

I put together what we jokingly called the "Jerry Maguire memo," which really functioned as a mission statement. The core idea was simple: digital is not just a channel. It's an enterprise-wide capability.

Today, we treat digital as its own business unit. It's a core driver of growth, member experience, revenue, and operational efficiency.

Chuck Hayes: What does it mean in practice to treat digital as a growth engine?

Ben Fetterman: It means digital isn't just about interfaces, it's about outcomes.

For us, digital drives how members engage with us anytime, anywhere. It impacts acquisition, retention, wallet share, and even cost structure.

We've made platform investments, re-evaluated vendors and brought in new capabilities. All of that is aligned with a broader goal: use digital to improve processes, enhance experiences, and enable scalable growth.

From a growth perspective, we've increased online account opening activity by 69% and doing so with 81% of it being automated compared to 0% prior. We've also driven strong growth trends in the traditional online and mobile banking channels averaging over 20% growth in active users and logins across our consumer and business channels.

Chuck Hayes: AI is a major focus across the industry. Where do you see the biggest misconceptions?

Ben Fetterman: The biggest issue is that AI is often treated as one thing, when in reality it's not.

People tend to lump conversational AI, predictive models, and agentic AI together, but they're very different in terms of how they function and what they require.

There's also a knowledge gap. Institutions understand that AI is important, but they don't always fully understand what's under the hood. Especially when it comes to data protection, governance, and implementation.

And there's pressure to move quickly, often driven by comparisons to non-financial companies. That can lead to implementations that aren't fully thought through, which creates risk and can negatively impact the member experience.

Chuck Hayes: How are you balancing innovation with member expectations?

Ben Fetterman: Members still expect choice.

We're not in a world where digital replaces everything else. Members want flexibility: branches, call centers, mobile, and online. AI fits into that ecosystem, but it doesn't eliminate the need for human interaction.

We've seen this pattern before. The technology evolves, but the need for high-touch service remains. The key is making sure the experience, regardless of channel, is consistent and effective.

Chuck Hayes: What was Trumark's first step into AI?

Ben Fetterman: We started with conversational AI about four years ago, when we launched a chatbot named "Mark."

The goal was to create a controlled environment where we could test AI safely: handling basic inquiries, reducing call volume, and providing 24/7 availability.

It worked well from an efficiency standpoint. It helped us build confidence internally and demonstrated that AI could deliver tangible value.

Chuck Hayes: At what point did your approach to AI evolve?

Ben Fetterman: We initially planned to expand conversational AI into our IVR (Interactive Voice Response) system. But after evaluating it, we decided to pause.

The concern was around member experience. The interaction risked feeling too robotic, and that's not what we wanted.

Instead, we made a conscious decision to shift our focus inward. As a smaller institution, we have the flexibility to be agile. We asked ourselves where AI could have the most immediate operational impact, and that led us to internal use cases.

Chuck Hayes: What types of internal use cases have you focused on?

Ben Fetterman: We've leaned heavily into agentic AI. AI that can actually execute tasks and manage processes end-to-end.

A strong example is loan quality control. Traditionally, reviewing a home equity loan could take around 80 minutes. With agentic AI, we've reduced that to under 10 minutes.

The AI reviews documentation, identifies missing items, and even generates communication back to the member to resolve gaps. That speeds up the entire process and reduces friction.

We've also applied AI to consumer loan underwriting and automated account opening workflows, where documents are routed directly into storage systems without manual intervention.

Chuck Hayes: How important is data in making these AI initiatives successful?

Ben Fetterman: It's critical. AI is only as good as the data behind it.

We created a dedicated data division and brought in leadership specifically focused on building a data strategy and governance. Before we scale AI further, specifically on the member side, we're building a data platform to centralize and manage our data, including building a governance plan, and addressing inconsistencies. Without that foundation, AI can actually amplify problems instead of solving them, especially when you get into member insights.

Chuck Hayes: How do CRM and data platforms fit into this strategy?

Ben Fetterman: We're implementing a combination of CRM, data platforms, and middleware to connect everything.

The goal is to create a unified data environment where information flows consistently across systems. Once that's in place, AI can be layered on top to enable things like real-time insights and more accurate and timely decision-making.

Chuck Hayes: Where do you see AI having the biggest impact once the data foundation is in place?

Ben Fetterman: One of the biggest opportunities is on the member-facing side with personalization.

With clean, structured data, we can run smarter, more personalized marketing campaigns that align with what our members want and need in their financial journey. This helps strengthen relevance, trust, and engagement.

Beyond campaign relevance, this gives us the ability to be proactive. We'll have the ability to recognize individual trends and even notify a member before something happens. Think about getting notified 10 days before you overdraw your account, based on your prior activity and expected bills, stating that you will be $150 short before your next paycheck. Insights like this are what creates those individualized/personalized experiences that our members expect these days.

AI allows us to tailor experiences in a way that feels timely and contextually appropriate.

Chuck Hayes: How does AI impact workforce planning and operational scalability?

Ben Fetterman: We're seeing a wave of experienced professionals retiring, taking institutional knowledge with them. At the same time, younger generations aren't necessarily drawn to repetitive, manual roles in operations.

AI helps bridge that gap.

By automating a large portion of tasks, up to 95% in some workflows, we can rely on a smaller number of experts to oversee and refine processes rather than execute everything manually.

This allows us to scale without proportionally increasing headcount.

Chuck Hayes: What do you see as the most underrated use cases for AI in financial services?

Ben Fetterman: The support office.

A lot of attention goes to front-end features, chatbots, digital assistants, and so on. Those are important, but the real impact often comes from improving internal workflows and processes, which in turn drives those front-end features.

When you make processes faster and more accurate behind the scenes, the member experience improves naturally. Loan processing, document handling, and internal workflows all benefit.

Agent assist is another underrated area. Giving employees real-time access to accurate information reduces friction in member interactions and improves consistency.

Chuck Hayes: From your perspective, where do fintech companies sometimes misunderstand credit unions?

Ben Fetterman: Fintechs often underestimate the time and process required for decision-making within credit unions.

We operate in a more collaborative environment. Decisions typically involve multiple stakeholders, internal alignment, and translation of technical concepts for non-technical audiences.

Fintechs are often very fast-moving and aggressive in closing deals, which doesn't always align with how credit unions evaluate and implement solutions.

The key is understanding that alignment and timing matter just as much as the technology itself.

Chuck Hayes: Any final thoughts on how institutions should approach AI?

Ben Fetterman: Start with clarity.

AI is powerful, but it works best when applied to a well-defined strategy and roadmap that are designed to solve both internal and external problems, enhance operational efficiency and improve the overall user experience.

The temptation is to move quickly because everyone is talking about AI. But the organizations that succeed will be the ones that focus on getting the fundamentals right first.

Final Thoughts

As financial institutions continue navigating the AI landscape, Trumark's journey offers a pragmatic perspective: prioritize the areas where AI can deliver the greatest impact with the least risk.

That doesn't mean ignoring member-facing innovation, Trumark has already seen value through conversational AI, but it does mean being intentional about where and how AI is deployed. For Trumark, the support office has proven to be the fastest path to meaningful gains, driving efficiency, accuracy, and scalability without introducing unnecessary exposure.

As the technology matures and data foundations strengthen, member-facing experiences will continue to evolve in parallel.

In a rapidly changing environment, that balance between innovation, risk, and operational readiness may ultimately define the institutions that win.